1. Suppose (unobserved) ability positively influences the number of years of schooling and wages. Explain how the absence of ability in the regression model above may lead to an overestimation of the impact of schooling on wages.

Suppose we observe data on hourly wages, wi the number of years of schooling, si and years of experience, xi, for person i, where i=1,2,,N. We can run a wage regression of the following: log(wi)=1+2si+3xi+4(xi)2+i, where i is the error term. Our interest is the parameter 2, which captures the relationship between years of schooling and wages, conditional on years of experience and its square. Suppose we observe data on hourly wages, wi the number of years of schooling, si and years of experience, xi, for person i, where i=1,2,,N. We can run a wage regression of the following: log(wi)=1+2si+3xi+4(xi)2+i, where i is the error term. Our interest is the parameter 2, which captures the relationship between years of schooling and wages, conditional on years of experience and its square